fix(bump): Search for version number line by line

Avoid the complexities of multiline regex matching and inconsistencies
with the case where no regex is specified. Simplify the implementation
by eliminating the need to track an offset into the file. Ensure that
the version number will be found even if it is to the left of the
portion of the line matching the regex.

Author: Kurt-von-Laven

commitizen/bump.py

@@ -142,18 +142,12 @@ def update_version_in_files(
     # TODO: separate check step and write step
     for location in files:
         filepath, _, regex = location.partition(":")
+        if not regex:
+            regex = _version_to_regex(current_version)
 
-        with open(filepath, "r") as f:
-            version_file = f.read()
-
-        if regex:
-            current_version_found, version_file = _bump_with_regex(
-                version_file, current_version, new_version, regex
-            )
-        else:
-            current_version_regex = _version_to_regex(current_version)
-            current_version_found = bool(current_version_regex.search(version_file))
-            version_file = current_version_regex.sub(new_version, version_file)
+        current_version_found, version_file = _bump_with_regex(
+            filepath, current_version, new_version, re.compile(regex)
+        )
 
         if check_consistency and not current_version_found:
             raise CurrentVersionNotFoundError(
@@ -168,33 +162,24 @@ def update_version_in_files(
 
 
 def _bump_with_regex(
-    version_file_contents: str, current_version: str, new_version: str, regex: str
+    version_filepath: str, current_version: str, new_version: str, regex: re.Pattern
 ) -> Tuple[bool, str]:
     current_version_found = False
-    # Bumping versions that change the string length move the offset on the file contents as finditer keeps a
-    # reference to the initial string that was used and calling search many times would lead in infinite loops
-    # e.g.: 1.1.9 -> 1.1.20
-    offset = 0
-    for match in re.finditer(regex, version_file_contents, re.MULTILINE):
-        left = version_file_contents[: match.end() + offset]
-        right = version_file_contents[match.end() + offset :]
-
-        line_break = right.find("\n")
-        middle = right[:line_break]
-        right = right[line_break:]
-
-        if current_version in middle:
-            offset += len(new_version) - len(current_version)
-            current_version_found = True
-            version_file_contents = (
-                left + middle.replace(current_version, new_version) + right
-            )
-    return current_version_found, version_file_contents
-
-
-def _version_to_regex(version: str) -> re.Pattern:
-    clean_regex = version.replace(".", r"\.").replace("+", r"\+")
-    return re.compile(f"{clean_regex}")
+    lines = []
+    with open(version_filepath, "r") as f:
+        for line in f:
+            if regex.search(line):
+                bumped_line = line.replace(current_version, new_version)
+                if bumped_line != line:
+                    current_version_found = True
+                lines.append(bumped_line)
+            else:
+                lines.append(line)
+    return current_version_found, "".join(lines)
+
+
+def _version_to_regex(version: str) -> str:
+    return version.replace(".", r"\.").replace("+", r"\+")
 
 
 def normalize_tag(